The challenges for data science arise because key entities are isolated from one another: methodology developers, methodology consumers, scientists, research software engineers.
A nice way to visualise isolated entities is as islands. And the key objective of open data science is to bridge between those islands. Islands represent users, data analysts and data scientists.
Think of the Florida Keys and the “Overseas Highway”
Just like the Overseas Highway, we can bridge better by building on the natural topology. A key component of that topology is the landscape of scientific software software specialists including research software engineers, developers of tools such as the jupyter notebook etc. These groups form the spine of our bridging.
The principle of the Open Data Science Initiative is that no one should be unwillingly isolated, but just like the Keys, we need to appreciate that each island is a delicate ecosystem on its own, and that our bridging should not be damaging what already exists, but enhancing it.
Open data science is not just about open source software, being exclusively open source can even be damaging to the agenda, because it prevents delivery of solutions to those who most need them. Even if we feel a particular software choice (such as Excel) is detrimental to long term understanding of data, it is important to bridge to its users to ensure that our solutions cater for their needs.
Our strategy will evolve, so this page will evolve with it. When the right strategy is unclear its important not to constrain things too much, but here we write some agreed upon ideas.